期刊名称：International Journal of Signal Processing, Image Processing and Pattern Recognition
印刷版ISSN：2005-4254
出版年度：2016
卷号：9
期号：9
页码：357-368
出版社：SERSC
摘要：Projectionmatrix plays an important role in compressive sensing(CS).Small mutual coherencebetween a projectionmatrixand asparsifying matrixis considered to enhance reconstructionperformancein CS.The equiangular tight frame(ETF) was demonstratedwith minimummutual coherence in previous works. However, ETF does not exist for any dimensions. A practical solution is to make the production of a projectionmatrixand a sparsifying matrix becomethe nearest one to ETF.Here, the optimizationobjective is regarded as the minimizationof error betweenthem. First, a maximum-likelihoodestimationmodel is presentedto crack the minimizationproblem.Thenan alternative multiplicativeiteration method is employed to guaranteethat theerror will convergeto the minimumefficiently.Experimentalresults showthat the proposed methodobtainssmaller mutual coherence withbetter reconstruction performancecompared to existing methods.
关键词：compressed sensing; ;projection;matrix; alternative multiplicative;average ;mutual coherence
